Figures & data
Figure 1 Scatter plot of PNI on the x-axis versus Martingale residuals (n=375). Patients above the horizontal line (zero) were at an increased risk for death compared with the expected risk from the Cox proportional hazards regression model. The curved line represents a smoother scatter plot. The point at which the smoother line crosses the horizontal line (zero) occurs at 42, indicating that this would be the best cut-off point to predict death from esophageal squamous cell carcinoma based on PNI. The point at which the curve line crosses the horizontal line (zero) is at a PNI of 52, indicating that this would be another cut-off point. With the help of the fit line on the scatter plot, we classified the patients into three categories according to PNI, ie, <42, 42–52, and >52.
Table 1 Baseline characteristics of 375 patients with ESCC
Table 2 Relationship between PNI and clinicopathological characteristics
Figure 2 Kaplan–Meier survival curves stratified by PNI.
Abbreviations: CSS, cancer-specific survival; PNI, prognostic nutritional index.
Table 3 Univariate and multivariate analyses in patients with ESCC